new power supply but still have some issues...update: northwood must be toast

[Duvie]

When I run prime 95 at voltages that worked 4-5 days ago and with temps never getting over 40c (program doesn't stay running long genough) I get these errors....

Also If I am running prime95 and I open up cadd this is the error I get....

Also I thik I notice a bit more explorer errors when doing something cpu intensive like dvd-divx....


This is all at 1.8@2.4ghz @1.675v...last week prime would run 12hours on 2.4ghz and 1.65v....never saw a temp higher then 52c on this machine....

Errors are in win98...I need to install winxp again...computer doesn't freeze just indivdual programs get errors and le me close them...like I said no freezing....


UPDATE:

I gets errors if I multitask anything....I tried each ram stick separately...I tried them at 4:3 ratio for 200mhz ddr...still errors!!!

I then drop fsb to 112fsb and run a 3:5 ratio for like 386mhz ddr...prime ran straight for an hour, about 10 times longer then it has been...how can it be memory???

I bumped it up to 120fsb and it ran again solid no errors...encoded a divx movie overnight...flawless...then try to run prime95...program error almost immediately...


To KenAF who said electromigration had nothing to do with voltage alone but with heat, obviously must have been pulling crap out of his arse...I have never seen a temp higher then 52c and that was for a brief period and in most apps my temps never go above 45c...Yet what was stable at 2.5ghz and 1.775v now can't get 2.2ghz stable at 1.725v...what was stable at 2.4ghz at 1.65v is no where close...

And it is diminishing I believe daily...what was stable last night is error plagued now...

I expect dead cpu or errors at default speed by next week....


here are a sample of 2 errors:

ACAD caused an invalid page fault in
module DAO350.DLL at 015f:044e2b75.
Registers:
EAX=00000000 CS=015f EIP=044e2b75 EFLGS=00010206
EBX=00000000 SS=0167 ESP=01e3fa04 EBP=01e3fb54
ECX=00000200 DS=0167 ESI=00c4a8d1 FS=138f
EDX=0000001e ES=0167 EDI=00000000 GS=0000
Bytes at CS:EIP:
8b 47 20 85 c0 74 08 3b c8 0f 87 16 01 00 00 8b
Stack dump:
01e3fb54 00c4abcd 00c4a8d1 00c4a90b 00000000 00000008 044d4523 00000000 00000200 01e3fb54 00c4abcd 00c4abcd 00000000 00000070 04477410 00000000


PRIME95 caused an invalid page fault in
module PRIME95.EXE at 015f:005c8358.
Registers:
EAX=000000bf CS=015f EIP=005c8358 EFLGS=00010282
EBX=00000000 SS=0167 ESP=01c9fc58 EBP=01f6a540
ECX=00000000 DS=0167 ESI=03c75b40 FS=31ef
EDX=0000006b ES=0167 EDI=01cb1000 GS=0000
Bytes at CS:EIP:
01 00 66 0f 28 ae 90 10 01 00 66 0f 28 b6 a0 10
Stack dump:
00466951 01c9fd3e 00000001 01c9fd5c 0041708b 00000050 00000000 00000008 00000003 00440400 00000000 00000190 b8f814c6 00000000 69a8e00c 00fc0001

prime95 error happened shortly after cadd error...both programs were running and explorer was being connected to the internet...


are these ram errors or cpu errors??? UPDATE: I think these are cache of cpu failing....

I tested the cpu and it encoded a dvd to divx in two pass mode at same speed it did 2 weeks ago when everything was fine...temnps never got over 45c then...


try running multiple appz and problems arise...


PLEASE HELP......

 

[RalfHutter]

I suggested that you run Memtest86 on your RAM. Did you ever run it?

 

[Duvie]

I am going to try, but read above on update....I think it shows it is not the ram....


I may also try to update the bios to 1005, but 1004c ran flawlessly for 3 weeks....


Those stack dumps seem like cpu memory addresses...don't they???

 

[Duvie]

tested ram with memtest..ran 10 passes of all tests including extended tests...no errors...

Things I think I need to do: (please tell me if you think these could be issues...)


1) remove hsf and reapply AS2 and cpu....temps are at 40c or below for most of the errors, can't believe heat can be an issue...

Is there a program that can test the l1 and l2 cache of the p4???I really think errors above are showing cpu cache addresses...

2) complete fdsk, repartition, reformat and only install win98 at this time...no dual boot

could a corrupted boot sector do this??? I have done a full reformat and winxp install and an over the top install of win98
...no reformat there...no fdsk either...

3) update board bios to 1005 from 1004c...if 1005 in official now..

 

[THUGSROOK]

all of the above 😉

 

[Jerboy]

The diagnostic process can be described in three words.

"Trial-and-error".

It's almost impossible to test a computer component without a special testing equipment. Just like it's mechanical and electrical counterparts, the parts may have been defective or there has been an underlying cause that destroyed a component, meaning your processor may be bad, but you could fry your new processor again if there is an underlying cause..

 

[WarCon]

Try reinstalling, would be my suggestion. (Install at default voltage, so you don't induce any errors into your situation to make you feel any worse.)

 

[Duvie]

I installed at default speed and low memory ratio on a newly partitioned and refoprmatted drive...still get errors....

I still am going to try new bios and reinstall cpu though temps look fine as they always have and mbm basically confirms temps when I ran it....


The real fear is like jerboy stated...what if it is the mobo ovevolting, damaging the cpu, or producing the errors....

I am going to pull cmos battery to make sure everything goes back to factory defaults and repull and lap cpu....

If this does not work I may just wait for its inevitable death and get another one...maybe buy a 1.6a before I get replacement....


NORTHWOOD owners beware...I think voltages regardless of heat is in fact dangerous....I have had much lower temps then most ppl have on this board...I just pressed it a bit too far I think at 1.775v...My program that was running when I first ran into the issue was lame mp3 in my gknot divx program...it got up to 45c max when I noticed it was running at low priority and taking way too long...I changed priority to normal and it sped up to right speed for cpu then shortly after I got a BSOD and slowly over the next week each drop down in speed became less and less stable...I mean 2.2ghz works one night in prime and entire divx encoding then can't run prime for 13 tests the next day....Hell I can't even get 2.0ghz to work at 1.75v...I used to get 2.4ghz at 1.65v...